2 weeks ago - last edited 2 weeks ago
i use rapidminer for sentiment analysis my modle is so simple but wheb it comes to the classifier it takes 2 days for processing i try svm - knn - neural net- only naive bayes takes about 2 hours which is two slow as i think
my dataset contain 12500 positive document and 12500 negative document of a movie reviews , also when i change the data set to another data set contain 1000 positive and 1000 negative it runs much more faster but it gives me a very bad accuracy ratio
please i search for days for this problem without getting any answer can any one help me
my code in xml attached
wating for any reply
2 weeks ago
Yes, that could very well happen. I haven't downloaded your XML because I'm late and running out the door but if you aren't doing pruning you should do it. Also, if you use n-grams, it will blow up the columns and make for longer processing.